blob: 5ff3a570cc81ed3ae5dc67f99c95105113b3f3af [file] [log] [blame]
// Copyright 2011 the V8 project authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
// Declares a Simulator for MIPS instructions if we are not generating a native
// MIPS binary. This Simulator allows us to run and debug MIPS code generation
// on regular desktop machines.
// V8 calls into generated code via the GeneratedCode wrapper,
// which will start execution in the Simulator or forwards to the real entry
// on a MIPS HW platform.
#ifndef V8_MIPS64_SIMULATOR_MIPS64_H_
#define V8_MIPS64_SIMULATOR_MIPS64_H_
// globals.h defines USE_SIMULATOR.
#include "src/common/globals.h"
#if defined(USE_SIMULATOR)
// Running with a simulator.
#include "src/base/hashmap.h"
#include "src/codegen/assembler.h"
#include "src/execution/simulator-base.h"
#include "src/mips64/constants-mips64.h"
#include "src/utils/allocation.h"
namespace v8 {
namespace internal {
// -----------------------------------------------------------------------------
// Utility functions
class CachePage {
public:
static const int LINE_VALID = 0;
static const int LINE_INVALID = 1;
static const int kPageShift = 12;
static const int kPageSize = 1 << kPageShift;
static const int kPageMask = kPageSize - 1;
static const int kLineShift = 2; // The cache line is only 4 bytes right now.
static const int kLineLength = 1 << kLineShift;
static const int kLineMask = kLineLength - 1;
CachePage() {
memset(&validity_map_, LINE_INVALID, sizeof(validity_map_));
}
char* ValidityByte(int offset) {
return &validity_map_[offset >> kLineShift];
}
char* CachedData(int offset) {
return &data_[offset];
}
private:
char data_[kPageSize]; // The cached data.
static const int kValidityMapSize = kPageSize >> kLineShift;
char validity_map_[kValidityMapSize]; // One byte per line.
};
class SimInstructionBase : public InstructionBase {
public:
Type InstructionType() const { return type_; }
inline Instruction* instr() const { return instr_; }
inline int32_t operand() const { return operand_; }
protected:
SimInstructionBase() : operand_(-1), instr_(nullptr), type_(kUnsupported) {}
explicit SimInstructionBase(Instruction* instr) {}
int32_t operand_;
Instruction* instr_;
Type type_;
private:
DISALLOW_ASSIGN(SimInstructionBase);
};
class SimInstruction : public InstructionGetters<SimInstructionBase> {
public:
SimInstruction() {}
explicit SimInstruction(Instruction* instr) { *this = instr; }
SimInstruction& operator=(Instruction* instr) {
operand_ = *reinterpret_cast<const int32_t*>(instr);
instr_ = instr;
type_ = InstructionBase::InstructionType();
DCHECK(reinterpret_cast<void*>(&operand_) == this);
return *this;
}
};
class Simulator : public SimulatorBase {
public:
friend class MipsDebugger;
// Registers are declared in order. See SMRL chapter 2.
enum Register {
no_reg = -1,
zero_reg = 0,
at,
v0, v1,
a0, a1, a2, a3, a4, a5, a6, a7,
t0, t1, t2, t3,
s0, s1, s2, s3, s4, s5, s6, s7,
t8, t9,
k0, k1,
gp,
sp,
s8,
ra,
// LO, HI, and pc.
LO,
HI,
pc, // pc must be the last register.
kNumSimuRegisters,
// aliases
fp = s8
};
// Coprocessor registers.
// Generated code will always use doubles. So we will only use even registers.
enum FPURegister {
f0, f1, f2, f3, f4, f5, f6, f7, f8, f9, f10, f11,
f12, f13, f14, f15, // f12 and f14 are arguments FPURegisters.
f16, f17, f18, f19, f20, f21, f22, f23, f24, f25,
f26, f27, f28, f29, f30, f31,
kNumFPURegisters
};
// MSA registers
enum MSARegister {
w0,
w1,
w2,
w3,
w4,
w5,
w6,
w7,
w8,
w9,
w10,
w11,
w12,
w13,
w14,
w15,
w16,
w17,
w18,
w19,
w20,
w21,
w22,
w23,
w24,
w25,
w26,
w27,
w28,
w29,
w30,
w31,
kNumMSARegisters
};
explicit Simulator(Isolate* isolate);
~Simulator();
// The currently executing Simulator instance. Potentially there can be one
// for each native thread.
V8_EXPORT_PRIVATE static Simulator* current(v8::internal::Isolate* isolate);
// Accessors for register state. Reading the pc value adheres to the MIPS
// architecture specification and is off by a 8 from the currently executing
// instruction.
void set_register(int reg, int64_t value);
void set_register_word(int reg, int32_t value);
void set_dw_register(int dreg, const int* dbl);
int64_t get_register(int reg) const;
double get_double_from_register_pair(int reg);
// Same for FPURegisters.
void set_fpu_register(int fpureg, int64_t value);
void set_fpu_register_word(int fpureg, int32_t value);
void set_fpu_register_hi_word(int fpureg, int32_t value);
void set_fpu_register_float(int fpureg, float value);
void set_fpu_register_double(int fpureg, double value);
void set_fpu_register_invalid_result64(float original, float rounded);
void set_fpu_register_invalid_result(float original, float rounded);
void set_fpu_register_word_invalid_result(float original, float rounded);
void set_fpu_register_invalid_result64(double original, double rounded);
void set_fpu_register_invalid_result(double original, double rounded);
void set_fpu_register_word_invalid_result(double original, double rounded);
int64_t get_fpu_register(int fpureg) const;
int32_t get_fpu_register_word(int fpureg) const;
int32_t get_fpu_register_signed_word(int fpureg) const;
int32_t get_fpu_register_hi_word(int fpureg) const;
float get_fpu_register_float(int fpureg) const;
double get_fpu_register_double(int fpureg) const;
template <typename T>
void get_msa_register(int wreg, T* value);
template <typename T>
void set_msa_register(int wreg, const T* value);
void set_fcsr_bit(uint32_t cc, bool value);
bool test_fcsr_bit(uint32_t cc);
bool set_fcsr_round_error(double original, double rounded);
bool set_fcsr_round64_error(double original, double rounded);
bool set_fcsr_round_error(float original, float rounded);
bool set_fcsr_round64_error(float original, float rounded);
void round_according_to_fcsr(double toRound, double& rounded,
int32_t& rounded_int, double fs);
void round64_according_to_fcsr(double toRound, double& rounded,
int64_t& rounded_int, double fs);
void round_according_to_fcsr(float toRound, float& rounded,
int32_t& rounded_int, float fs);
void round64_according_to_fcsr(float toRound, float& rounded,
int64_t& rounded_int, float fs);
template <typename T_fp, typename T_int>
void round_according_to_msacsr(T_fp toRound, T_fp& rounded,
T_int& rounded_int);
void set_fcsr_rounding_mode(FPURoundingMode mode);
void set_msacsr_rounding_mode(FPURoundingMode mode);
unsigned int get_fcsr_rounding_mode();
unsigned int get_msacsr_rounding_mode();
// Special case of set_register and get_register to access the raw PC value.
void set_pc(int64_t value);
int64_t get_pc() const;
Address get_sp() const { return static_cast<Address>(get_register(sp)); }
// Accessor to the internal simulator stack area.
uintptr_t StackLimit(uintptr_t c_limit) const;
// Executes MIPS instructions until the PC reaches end_sim_pc.
void Execute();
template <typename Return, typename... Args>
Return Call(Address entry, Args... args) {
return VariadicCall<Return>(this, &Simulator::CallImpl, entry, args...);
}
// Alternative: call a 2-argument double function.
double CallFP(Address entry, double d0, double d1);
// Push an address onto the JS stack.
uintptr_t PushAddress(uintptr_t address);
// Pop an address from the JS stack.
uintptr_t PopAddress();
// Debugger input.
void set_last_debugger_input(char* input);
char* last_debugger_input() { return last_debugger_input_; }
// Redirection support.
static void SetRedirectInstruction(Instruction* instruction);
// ICache checking.
static bool ICacheMatch(void* one, void* two);
static void FlushICache(base::CustomMatcherHashMap* i_cache, void* start,
size_t size);
// Returns true if pc register contains one of the 'special_values' defined
// below (bad_ra, end_sim_pc).
bool has_bad_pc() const;
private:
enum special_values {
// Known bad pc value to ensure that the simulator does not execute
// without being properly setup.
bad_ra = -1,
// A pc value used to signal the simulator to stop execution. Generally
// the ra is set to this value on transition from native C code to
// simulated execution, so that the simulator can "return" to the native
// C code.
end_sim_pc = -2,
// Unpredictable value.
Unpredictable = 0xbadbeaf
};
V8_EXPORT_PRIVATE intptr_t CallImpl(Address entry, int argument_count,
const intptr_t* arguments);
// Unsupported instructions use Format to print an error and stop execution.
void Format(Instruction* instr, const char* format);
// Helpers for data value tracing.
enum TraceType {
BYTE,
HALF,
WORD,
DWORD,
FLOAT,
DOUBLE,
FLOAT_DOUBLE,
WORD_DWORD
};
// MSA Data Format
enum MSADataFormat { MSA_VECT = 0, MSA_BYTE, MSA_HALF, MSA_WORD, MSA_DWORD };
typedef union {
int8_t b[kMSALanesByte];
uint8_t ub[kMSALanesByte];
int16_t h[kMSALanesHalf];
uint16_t uh[kMSALanesHalf];
int32_t w[kMSALanesWord];
uint32_t uw[kMSALanesWord];
int64_t d[kMSALanesDword];
uint64_t ud[kMSALanesDword];
} msa_reg_t;
// Read and write memory.
inline uint32_t ReadBU(int64_t addr);
inline int32_t ReadB(int64_t addr);
inline void WriteB(int64_t addr, uint8_t value);
inline void WriteB(int64_t addr, int8_t value);
inline uint16_t ReadHU(int64_t addr, Instruction* instr);
inline int16_t ReadH(int64_t addr, Instruction* instr);
// Note: Overloaded on the sign of the value.
inline void WriteH(int64_t addr, uint16_t value, Instruction* instr);
inline void WriteH(int64_t addr, int16_t value, Instruction* instr);
inline uint32_t ReadWU(int64_t addr, Instruction* instr);
inline int32_t ReadW(int64_t addr, Instruction* instr, TraceType t = WORD);
inline void WriteW(int64_t addr, int32_t value, Instruction* instr);
void WriteConditionalW(int64_t addr, int32_t value, Instruction* instr,
int32_t rt_reg);
inline int64_t Read2W(int64_t addr, Instruction* instr);
inline void Write2W(int64_t addr, int64_t value, Instruction* instr);
inline void WriteConditional2W(int64_t addr, int64_t value,
Instruction* instr, int32_t rt_reg);
inline double ReadD(int64_t addr, Instruction* instr);
inline void WriteD(int64_t addr, double value, Instruction* instr);
template <typename T>
T ReadMem(int64_t addr, Instruction* instr);
template <typename T>
void WriteMem(int64_t addr, T value, Instruction* instr);
// Helper for debugging memory access.
inline void DieOrDebug();
void TraceRegWr(int64_t value, TraceType t = DWORD);
template <typename T>
void TraceMSARegWr(T* value, TraceType t);
template <typename T>
void TraceMSARegWr(T* value);
void TraceMemWr(int64_t addr, int64_t value, TraceType t);
void TraceMemRd(int64_t addr, int64_t value, TraceType t = DWORD);
template <typename T>
void TraceMemRd(int64_t addr, T value);
template <typename T>
void TraceMemWr(int64_t addr, T value);
// Operations depending on endianness.
// Get Double Higher / Lower word.
inline int32_t GetDoubleHIW(double* addr);
inline int32_t GetDoubleLOW(double* addr);
// Set Double Higher / Lower word.
inline int32_t SetDoubleHIW(double* addr);
inline int32_t SetDoubleLOW(double* addr);
SimInstruction instr_;
// functions called from DecodeTypeRegister.
void DecodeTypeRegisterCOP1();
void DecodeTypeRegisterCOP1X();
void DecodeTypeRegisterSPECIAL();
void DecodeTypeRegisterSPECIAL2();
void DecodeTypeRegisterSPECIAL3();
void DecodeTypeRegisterSRsType();
void DecodeTypeRegisterDRsType();
void DecodeTypeRegisterWRsType();
void DecodeTypeRegisterLRsType();
int DecodeMsaDataFormat();
void DecodeTypeMsaI8();
void DecodeTypeMsaI5();
void DecodeTypeMsaI10();
void DecodeTypeMsaELM();
void DecodeTypeMsaBIT();
void DecodeTypeMsaMI10();
void DecodeTypeMsa3R();
void DecodeTypeMsa3RF();
void DecodeTypeMsaVec();
void DecodeTypeMsa2R();
void DecodeTypeMsa2RF();
template <typename T>
T MsaI5InstrHelper(uint32_t opcode, T ws, int32_t i5);
template <typename T>
T MsaBitInstrHelper(uint32_t opcode, T wd, T ws, int32_t m);
template <typename T>
T Msa3RInstrHelper(uint32_t opcode, T wd, T ws, T wt);
// Executing is handled based on the instruction type.
void DecodeTypeRegister();
inline int32_t rs_reg() const { return instr_.RsValue(); }
inline int64_t rs() const { return get_register(rs_reg()); }
inline uint64_t rs_u() const {
return static_cast<uint64_t>(get_register(rs_reg()));
}
inline int32_t rt_reg() const { return instr_.RtValue(); }
inline int64_t rt() const { return get_register(rt_reg()); }
inline uint64_t rt_u() const {
return static_cast<uint64_t>(get_register(rt_reg()));
}
inline int32_t rd_reg() const { return instr_.RdValue(); }
inline int32_t fr_reg() const { return instr_.FrValue(); }
inline int32_t fs_reg() const { return instr_.FsValue(); }
inline int32_t ft_reg() const { return instr_.FtValue(); }
inline int32_t fd_reg() const { return instr_.FdValue(); }
inline int32_t sa() const { return instr_.SaValue(); }
inline int32_t lsa_sa() const { return instr_.LsaSaValue(); }
inline int32_t ws_reg() const { return instr_.WsValue(); }
inline int32_t wt_reg() const { return instr_.WtValue(); }
inline int32_t wd_reg() const { return instr_.WdValue(); }
inline void SetResult(const int32_t rd_reg, const int64_t alu_out) {
set_register(rd_reg, alu_out);
TraceRegWr(alu_out);
}
inline void SetFPUWordResult(int32_t fd_reg, int32_t alu_out) {
set_fpu_register_word(fd_reg, alu_out);
TraceRegWr(get_fpu_register(fd_reg), WORD);
}
inline void SetFPUWordResult2(int32_t fd_reg, int32_t alu_out) {
set_fpu_register_word(fd_reg, alu_out);
TraceRegWr(get_fpu_register(fd_reg));
}
inline void SetFPUResult(int32_t fd_reg, int64_t alu_out) {
set_fpu_register(fd_reg, alu_out);
TraceRegWr(get_fpu_register(fd_reg));
}
inline void SetFPUResult2(int32_t fd_reg, int64_t alu_out) {
set_fpu_register(fd_reg, alu_out);
TraceRegWr(get_fpu_register(fd_reg), DOUBLE);
}
inline void SetFPUFloatResult(int32_t fd_reg, float alu_out) {
set_fpu_register_float(fd_reg, alu_out);
TraceRegWr(get_fpu_register(fd_reg), FLOAT);
}
inline void SetFPUDoubleResult(int32_t fd_reg, double alu_out) {
set_fpu_register_double(fd_reg, alu_out);
TraceRegWr(get_fpu_register(fd_reg), DOUBLE);
}
void DecodeTypeImmediate();
void DecodeTypeJump();
// Used for breakpoints and traps.
void SoftwareInterrupt();
// Compact branch guard.
void CheckForbiddenSlot(int64_t current_pc) {
Instruction* instr_after_compact_branch =
reinterpret_cast<Instruction*>(current_pc + kInstrSize);
if (instr_after_compact_branch->IsForbiddenAfterBranch()) {
FATAL(
"Error: Unexpected instruction 0x%08x immediately after a "
"compact branch instruction.",
*reinterpret_cast<uint32_t*>(instr_after_compact_branch));
}
}
// Stop helper functions.
bool IsWatchpoint(uint64_t code);
void PrintWatchpoint(uint64_t code);
void HandleStop(uint64_t code, Instruction* instr);
bool IsStopInstruction(Instruction* instr);
bool IsEnabledStop(uint64_t code);
void EnableStop(uint64_t code);
void DisableStop(uint64_t code);
void IncreaseStopCounter(uint64_t code);
void PrintStopInfo(uint64_t code);
// Executes one instruction.
void InstructionDecode(Instruction* instr);
// Execute one instruction placed in a branch delay slot.
void BranchDelayInstructionDecode(Instruction* instr) {
if (instr->InstructionBits() == nopInstr) {
// Short-cut generic nop instructions. They are always valid and they
// never change the simulator state.
return;
}
if (instr->IsForbiddenAfterBranch()) {
FATAL("Eror:Unexpected %i opcode in a branch delay slot.",
instr->OpcodeValue());
}
InstructionDecode(instr);
SNPrintF(trace_buf_, " ");
}
// ICache.
static void CheckICache(base::CustomMatcherHashMap* i_cache,
Instruction* instr);
static void FlushOnePage(base::CustomMatcherHashMap* i_cache, intptr_t start,
size_t size);
static CachePage* GetCachePage(base::CustomMatcherHashMap* i_cache,
void* page);
enum Exception {
none,
kIntegerOverflow,
kIntegerUnderflow,
kDivideByZero,
kNumExceptions
};
// Exceptions.
void SignalException(Exception e);
// Handle arguments and return value for runtime FP functions.
void GetFpArgs(double* x, double* y, int32_t* z);
void SetFpResult(const double& result);
void CallInternal(Address entry);
// Architecture state.
// Registers.
int64_t registers_[kNumSimuRegisters];
// Coprocessor Registers.
// Note: FPUregisters_[] array is increased to 64 * 8B = 32 * 16B in
// order to support MSA registers
int64_t FPUregisters_[kNumFPURegisters * 2];
// FPU control register.
uint32_t FCSR_;
// MSA control register.
uint32_t MSACSR_;
// Simulator support.
// Allocate 1MB for stack.
size_t stack_size_;
char* stack_;
bool pc_modified_;
int64_t icount_;
int break_count_;
EmbeddedVector<char, 128> trace_buf_;
// Debugger input.
char* last_debugger_input_;
v8::internal::Isolate* isolate_;
// Registered breakpoints.
Instruction* break_pc_;
Instr break_instr_;
// Stop is disabled if bit 31 is set.
static const uint32_t kStopDisabledBit = 1 << 31;
// A stop is enabled, meaning the simulator will stop when meeting the
// instruction, if bit 31 of watched_stops_[code].count is unset.
// The value watched_stops_[code].count & ~(1 << 31) indicates how many times
// the breakpoint was hit or gone through.
struct StopCountAndDesc {
uint32_t count;
char* desc;
};
StopCountAndDesc watched_stops_[kMaxStopCode + 1];
// Synchronization primitives.
enum class MonitorAccess {
Open,
RMW,
};
enum class TransactionSize {
None = 0,
Word = 4,
DoubleWord = 8,
};
// The least-significant bits of the address are ignored. The number of bits
// is implementation-defined, between 3 and minimum page size.
static const uintptr_t kExclusiveTaggedAddrMask = ~((1 << 3) - 1);
class LocalMonitor {
public:
LocalMonitor();
// These functions manage the state machine for the local monitor, but do
// not actually perform loads and stores. NotifyStoreConditional only
// returns true if the store conditional is allowed; the global monitor will
// still have to be checked to see whether the memory should be updated.
void NotifyLoad();
void NotifyLoadLinked(uintptr_t addr, TransactionSize size);
void NotifyStore();
bool NotifyStoreConditional(uintptr_t addr, TransactionSize size);
private:
void Clear();
MonitorAccess access_state_;
uintptr_t tagged_addr_;
TransactionSize size_;
};
class GlobalMonitor {
public:
class LinkedAddress {
public:
LinkedAddress();
private:
friend class GlobalMonitor;
// These functions manage the state machine for the global monitor, but do
// not actually perform loads and stores.
void Clear_Locked();
void NotifyLoadLinked_Locked(uintptr_t addr);
void NotifyStore_Locked();
bool NotifyStoreConditional_Locked(uintptr_t addr,
bool is_requesting_thread);
MonitorAccess access_state_;
uintptr_t tagged_addr_;
LinkedAddress* next_;
LinkedAddress* prev_;
// A scd can fail due to background cache evictions. Rather than
// simulating this, we'll just occasionally introduce cases where an
// store conditional fails. This will happen once after every
// kMaxFailureCounter exclusive stores.
static const int kMaxFailureCounter = 5;
int failure_counter_;
};
// Exposed so it can be accessed by Simulator::{Read,Write}Ex*.
base::Mutex mutex;
void NotifyLoadLinked_Locked(uintptr_t addr, LinkedAddress* linked_address);
void NotifyStore_Locked(LinkedAddress* linked_address);
bool NotifyStoreConditional_Locked(uintptr_t addr,
LinkedAddress* linked_address);
// Called when the simulator is destroyed.
void RemoveLinkedAddress(LinkedAddress* linked_address);
static GlobalMonitor* Get();
private:
// Private constructor. Call {GlobalMonitor::Get()} to get the singleton.
GlobalMonitor() = default;
friend class base::LeakyObject<GlobalMonitor>;
bool IsProcessorInLinkedList_Locked(LinkedAddress* linked_address) const;
void PrependProcessor_Locked(LinkedAddress* linked_address);
LinkedAddress* head_ = nullptr;
};
LocalMonitor local_monitor_;
GlobalMonitor::LinkedAddress global_monitor_thread_;
};
} // namespace internal
} // namespace v8
#endif // defined(USE_SIMULATOR)
#endif // V8_MIPS64_SIMULATOR_MIPS64_H_